Title: Vector Search: Fundamentals and Applications

Abstract: Vector search has become a foundational primitive in modern data systems, powering semantic retrieval, recommendation, and retrieval-augmented generation at scale. This tutorial covers the full stack. We begin with embeddings and explain why traditional indexes fail in high dimensional spaces. We then survey the four major index families (hash based, tree-based, graph-based, and quantization-based) and examine the unsolved problem of filtered vector search. We show how classical data engineering concerns such as freshness, provenance, schema migration, and data quality resurface in harder form when the data type is a vector.
Finally, we survey the emerging system landscape, from native vector databases to object-storage-native architectures, and close with a map of open research problems at the frontier of the field.